Complete-linkage clustering for voice activity detection in audio and visual speech

We propose a novel technique for conducting robust voice activity detection (VAD) in high-noise recordings. We use Gaussian mixture modeling (GMM) to train two generic models; speech and non-speech. We then score smaller segments of a given (unseen) recording against each of these GMMs to obtain two respective likelihood scores for each segment. These scores are used to compute a dissimilarity measure between pairs of segments and to carry out complete-linkage clustering of the segments into speech and non-speech clusters. We compare the accuracy of our method against state-of-the-art and standardised VAD techniques to demonstrate an absolute improvement of 15% in half-total error rate (HTER) over the best performing baseline system and across the QUT-NOISE-TIMIT database. We then apply our approach to the Audio-Visual Database of American English (AVDBAE) to demonstrate the performance of our algorithm in using visual, audio-visual or a proposed fusion of these features.

Art museums and the global tourist: Experience centres in experience-scapes

In the experience economy, the role of art museums has evolved so as to cater to global cultural tourists. These institutions were traditionally dedicated to didactic functions, and served cognoscenti with elite cultural tastes that were aligned with the avant-garde’s autonomous stance towards mass culture. In a post-avant-garde era however museums have focused on appealing to a broad clientele that often has little or no knowledge of historical or contemporary art. Many of these tourists want art to provide entertaining and novel experiences, rather than receiving pedagogical ‘training’. In response, art museums are turning into ‘experience venues’ and are being informed by ideas associated with new museology, as well as business approaches like Customer Experience Management. This has led to the provision of populist entertainment modes, such as blockbuster exhibitions, participatory art events, jazz nights, and wine tasting, and reveals that such museums recognize that today’s cultural tourist is part of an increasingly diverse and populous demographic, which shares many languages and value systems. As art museums have shifted attention to global tourists, they have come to play a greater role in gentrification projects and cultural precincts. The art museum now seems ideally suited to tourist-centric environments that offer a variety of immersive sensory experiences and combine museums (often designed by star-architects), international hotels, restaurants, high-end shopping zones, and other leisure forums. These include sites such as Port Maravilha urban waterfront development in Rio de Janiero, the Museum of Old and New Art in Hobart, and the Chateau La Coste winery and hotel complex in Provence. It can be argued that in a global experience economy, art museums have become experience centres in experience-scapes. This paper will examine the nature of the tourist experience in relation to the new art museum, and the latter’s increasingly important role in attracting tourists to urban and regional cultural precincts.

The QUT-NOISE-SRE protocol for the evaluation of noisy speaker recognition

The QUT-NOISE-SRE protocol is designed to mix the large QUT-NOISE database, consisting of over 10 hours of back- ground noise, collected across 10 unique locations covering 5 common noise scenarios, with commonly used speaker recognition datasets such as Switchboard, Mixer and the speaker recognition evaluation (SRE) datasets provided by NIST. By allowing common, clean, speech corpora to be mixed with a wide variety of noise conditions, environmental reverberant responses, and signal-to-noise ratios, this protocol provides a solid basis for the development, evaluation and benchmarking of robust speaker recognition algorithms, and is freely available to download alongside the QUT-NOISE database. In this work, we use the QUT-NOISE-SRE protocol to evaluate a state-of-the-art PLDA i-vector speaker recognition system, demonstrating the importance of designing voice-activity-detection front-ends specifically for speaker recognition, rather than aiming for perfect coherence with the true speech/non-speech boundaries.

Dynamics of reverse salience as technological performance gap: An empirical study of the personal computer technology system

The evolution of technological systems is hindered by systemic components, referred to as reverse salients, which fail to deliver the necessary level of technological performance thereby inhibiting the performance delivery of the system as a whole. This paper develops a performance gap measure of reverse salience and applies this measurement in the study of the PC (personal computer) technological system, focusing on the evolutions of firstly the CPU (central processing unit) and PC game sub-systems, and secondly the GPU (graphics processing unit) and PC game sub-systems. The measurement of the temporal behavior of reverse salience indicates that the PC game sub-system is the reverse salient, continuously trailing behind the technological performance of the CPU and GPU sub-systems from 1996 through 2006. The technological performance of the PC game sub-system as a reverse salient trails that of the CPU sub-system by up to 2300 MHz with a gradually decreasing performance disparity in recent years. In contrast, the dynamics of the PC game sub-system as a reverse salient trails the GPU sub-system with an ever increasing performance gap throughout the timeframe of analysis. In addition, we further discuss the research and managerial implications of our findings.

External knowledge and query strategies in active learning: A study in clinical information extraction

This paper presents a new active learning query strategy for information extraction, called Domain Knowledge Informativeness (DKI). Active learning is often used to reduce the amount of annotation effort required to obtain training data for machine learning algorithms. A key component of an active learning approach is the query strategy, which is used to iteratively select samples for annotation. Knowledge resources have been used in information extraction as a means to derive additional features for sample representation. DKI is, however, the first query strategy that exploits such resources to inform sample selection. To evaluate the merits of DKI, in particular with respect to the reduction in annotation effort that the new query strategy allows to achieve, we conduct a comprehensive empirical comparison of active learning query strategies for information extraction within the clinical domain. The clinical domain was chosen for this work because of the availability of extensive structured knowledge resources which have often been exploited for feature generation. In addition, the clinical domain offers a compelling use case for active learning because of the necessary high costs and hurdles associated with obtaining annotations in this domain. Our experimental findings demonstrated that 1) amongst existing query strategies, the ones based on the classification model’s confidence are a better choice for clinical data as they perform equally well with a much lighter computational load, and 2) significant reductions in annotation effort are achievable by exploiting knowledge resources within active learning query strategies, with up to 14% less tokens and concepts to manually annotate than with state-of-the-art query strategies.

IdeaWall: Mixed mode distributed collaboration in enterprise environments

In early stages of design and modeling, computers and computer applications are often considered an obstacle, rather than a facilitator of the process. Most notably, brainstorms, process modeling with business experts, or development planning, are often performed by a team in front of a whiteboard. While "whiteboarding" is recognized as an effective tool, low-tech solutions that allow remote participants to contribute are still not generally available. This is a striking observation, considering that vast majority of teams in large organizations are distributed teams. And this has also been one of the key triggers behind the project described in this article, where a team of corporate researchers decided to identify state of the art technologies that could facilitate the scenario mentioned above. This paper is an account of a research project in the area of enterprise collaboration, with a strong focus on the aspects of human computer interaction in mixed mode environments, especially in areas of collaboration where computers still play a secondary role. It is describing a currently running corporate research project. © 2012 Springer-Verlag.

IdeaWall: Bridging the digital and non-digital worlds to facilitate distant collaboration

Creative and ad-hoc work often involves non-digital artifacts, such as whiteboards and post-it notes. The preferred method of brainstorming and idea development, while facilitating work among collocated participants, makes it particularly tricky to involve remote participants, not even mentioning cases where live social involvement is required and the number and location of remote participants can be vast. Our work has originally focused on large distributed teams in business entities. Vast majority of teams in large organizations are distributed teams. Our team of corporate researchers decided to identify state of the art technologies that could facilitate the scenarios mentioned above. This paper is an account of a research project in the area of enterprise collaboration, with a strong focus on the aspects of human computer interaction in mixed mode environments, especially in areas of collaboration where computers still play a secondary role. It is describing a currently running corporate research project. In this paper we signal the potential use of the technology in situation, where community involvement is either required or desirable. The goal of the paper is to initiate a discussion on the use of technologies, initially designed as supporting enterprise collaboration, in situation requiring community engagement. In other words, it is a contribution of technically focused research exploring the uses of the technology in areas such as social engagement and community involvement. © 2012 IEEE.

Cross database training of audio-visual hidden Markov models for phone recognition

Speech recognition can be improved by using visual information in the form of lip movements of the speaker in addition to audio information. To date, state-of-the-art techniques for audio-visual speech recognition continue to use audio and visual data of the same database for training their models. In this paper, we present a new approach to make use of one modality of an external dataset in addition to a given audio-visual dataset. By so doing, it is possible to create more powerful models from other extensive audio-only databases and adapt them on our comparatively smaller multi-stream databases. Results show that the presented approach outperforms the widely adopted synchronous hidden Markov models (HMM) trained jointly on audio and visual data of a given audio-visual database for phone recognition by 29% relative. It also outperforms the external audio models trained on extensive external audio datasets and also internal audio models by 5.5% and 46% relative respectively. We also show that the proposed approach is beneficial in noisy environments where the audio source is affected by the environmental noise.

Learned ultra-wideband RADAR sensor model for augmented LIDAR-based traversability mapping in vegetated environments

In vegetated environments, reliable obstacle detection remains a challenge for state-of-the-art methods, which are usually based on geometrical representations of the environment built from LIDAR and/or visual data. In many cases, in practice field robots could safely traverse through vegetation, thereby avoiding costly detours. However, it is often mistakenly interpreted as an obstacle. Classifying vegetation is insufficient since there might be an obstacle hidden behind or within it. Some Ultra-wide band (UWB) radars can penetrate through vegetation to help distinguish actual obstacles from obstacle-free vegetation. However, these sensors provide noisy and low-accuracy data. Therefore, in this work we address the problem of reliable traversability estimation in vegetation by augmenting LIDAR-based traversability mapping with UWB radar data. A sensor model is learned from experimental data using a support vector machine to convert the radar data into occupancy probabilities. These are then fused with LIDAR-based traversability data. The resulting augmented traversability maps capture the fine resolution of LIDAR-based maps but clear safely traversable foliage from being interpreted as obstacle. We validate the approach experimentally using sensors mounted on two different mobile robots, navigating in two different environments.

Computer analysis of impact behavior of concrete filled steel tube columns

This paper presents a numerical study of the response of axially loaded concrete filled steel tube (CFST) columns under lateral impact loading using explicit non-linear finite element techniques. The aims of this paper are to evaluate the vulnerability of existing columns to credible impact events as well as to contribute new information towards the safe design of such vulnerable columns. The model incorporates concrete confinement, strain rate effects of steel and concrete, contact between the steel tube and concrete and dynamic relaxation for pre-loading, which is a relatively recent method for applying a pre-loading in the explicit solver. The finite element model was first verified by comparing results with existing experimental results and then employed to conduct a parametric sensitivity analysis. The effects of various structural and load parameters on the impact response of the CFST column were evaluated to identify the key controlling factors. Overall, the major parameters which influence the impact response of the column are the steel tube thickness to diameter ratio, the slenderness ratio and the impact velocity. The findings of this study will enhance the current state of knowledge in this area and can serve as a benchmark reference for future analysis and design of CFST columns under lateral impact.

Stone baby: An exploration of trembling and its potential for visual art

Informed by Kristeva's formulation of affect and Winnicott's Holding Environment, this practice-led visual art project is an exploration into how sensitivity to the physical sensation of trembling can sustain a creative practice. Building upon this is a further enquiry into what the significance of the affective experience of trembling is for an ethics of affect in contemporary art. I have done this through object and video-based installations informed by my own experience of trembling. This has been further informed by the work of artists like Louise Bourgeois, Dennis Del Favero and Willie Doherty. The creative outcomes contribute to the discourse around ethical responses to affect by extending and developing on the works of these artists.

The politics of art/Art of politics

Author Toni Morrison said, “All good art is political! There is none that isn’t”. Perhaps this is why the arts and artists throughout history have been positioned as dangerous, troubling and on the margin. Art works can ask questions of us, challenge assumptions and name the un-nameable. Art works challenge hegemonies and the status quo – they trouble politics. So what happens when arts meets politics when it comes to the entitlement for young Australians to an arts-rich education? How do we navigate the tricky waters of the political ebb and flow to champion the agenda for arts education in contemporary classrooms so that our young people can be cultural navigators, cultural auteurs and culture makers?

New microscale vertically oriented organic photovoltaics cells

We initially look at the changing energy environment and how that can have a dramatic change on the potential of alternative energies, in particular those of organic photovoltaicvs (OPV) cells. In looking at OPV's we also address the aspects of where we are with the current art and why we may not be getting the best from our materials. In doing so, we propose the idea of changing how we build organic photovoltaics by addressing the best method to contain light within the devices. Our initial effort is in addressing how these microscale optical concentrators work in the form of optical fibers in terms of absorption. We have derived a mathematical method which takes account of the input angle of light to achieve optimum absorption. However, in doing so we also address the complex issue how the changing refractive indices in a multilayer device can alter how we input the light. We have found that by knowing the materials refractive index our model takes into account the incident plane, meridonal plane, cross sectional are and path length to ensure optical angular input. Secondly, we also address the practicalities of making such vertical structures the greater issue of changing light intensity incident on a solar cell and how that aspects alters how we view the performance of organic solar cells.

A right of resale? Indigenous art under the hammer

This week, Sotheby's sold the late Clifford Possum Tjapaltjarri's painting, Warlugulong. The auction-house's spokesman, Tim Klingender, was enthusiastic about the high price commanded by the art work: "The painting was a really great painting and it deserved to make a really fantastic price, and it made that price."